Precision Breeding and Consumer Safety: A Review of Regulations for UK Markets

Abstract

Gene-edited crops and livestock have the potential to transform food systems by providing resilience to climate change, pest and disease resistance, and the enhancement of nutrients in feed and food in a time-efficient and precise way. In 2023, the UK Parliament passed the Genetic Technology (Precision Breeding) Bill, paving the way for gene-edited products to be farmed in England and sold, providing they could have theoretically been produced via traditional breeding. In this paper, we describe the possible risks of gene-edited products for consumption using four case studies of gene-edited organisms: increased vitamin D tomatoes, reduced linoleic acid cottonseed oil, porcine reproductive and respiratory virus (PRRSV) resistant pigs and reduced-asparagine wheat. Assuming that the only requirement for an organism to be a Precision-Bred Organism (PBO) is that no transgenic material remains within the organism and that the edit could have, in theory, occurred spontaneously or through traditional breeding methods, then all our case studies would likely be defined as PBOs. We also conclude that the food safety risks of these products appear to be similar to those that society accepts in traditionally bred organisms used for food and feed. However, PBOs that possess markedly altered nutrient profiles may require a dedicated identity-preserved retail chain and/or labelling to avoid unintended over-consumption.

1. Introduction

The increasing global population, climate change, constraints on land and declining soil fertility are challenges for the agricultural sector [1]. It has been estimated that between 2010 and 2050, food production needs to increase by 32%, and radical changes in diet and food habits need to be made to feed the population and meet the United Nations’ Sustainable Development Goals [2,3]. Food waste is a major burden globally with an estimated 1/3 of food produced being lost or wasted; improving the shelf-life, stability and durability of products may decrease this [4]. Furthermore, micronutrient deficiencies, over-nourishment and malnourishment are global health threats that could be aided by technologies that increase the nutritional content of foods [5]. The use of biotechnology, in particular gene editing, may have the potential to tackle these global challenges. Precision breeding, which generates genetic changes that could have occurred spontaneously or via traditional processes of breeding, offers a precise and time-efficient way to tackle these issues and produce crops and livestock better suited to the current and future challenges of sustainable agriculture.

Precision breeding techniques, such as gene editing, offer an elegant way to produce crops and animals with specific traits that may otherwise be resource-intensive, unpredictable, or slow to produce through conventional breeding. Most gene-editing strategies utilise site-directed nucleases to create alterations at pre-determined points in an organism’s DNA [6]. Commonly used nucleases include Clustered Regularly Interspaced Short Palindromic Repeats (CRISPR) nucleases, transcription activator-like effector nucleases (TALENs) and zinc-finger nucleases [7]. The most widely adopted system, CRISPR/Cas9, consists of a single-guide RNA (sgRNA) coupled with endonucleases which cleaves the DNA, creating a double-stranded break [8]. These breaks are predominantly repaired by the cell via canonical non-homologous end joining, resulting in insertions or deletions of the target DNA [9]. Commercially grown products of gene editing such as high oleic soybean oil, drought-resistant corn, GABA-enriched tomatoes and fungus resistance canola have demonstrated the success of this technique [10].

Precision-bred organisms (PBOs) possess mutations including deletions, inversions, substitutions or additional DNA fragments that could have been obtained through conventional breeding approaches and where no foreign DNA remains in the resulting organism [11]. These techniques differ from those used to generate Genetically Modified Organisms (GMOs), where the organism’s genome has been altered in a way that would not naturally occur through mating or natural recombination and are generated via transgenesis. PBO legislation also allows cisgenesis, a process that transfers DNA between sexually compatible species outside of conventional reproductive processes.

A global shift towards the use of gene editing over the last decade has resulted in widespread adaptation of legislation for novel products to account for innovation in genetic technologies [11]. Led by early adopters including Argentina [12], many countries, with the exception of a handful such as those in the European Union (EU) and New Zealand, differentiate some products of gene editing from genetically modified organisms (GMOs) and, on a case-by-case basis, deem them safe for cultivation, import and consumption [13]. However, the regulatory landscape that assesses the associated risks differ markedly from country to country [14]. Some alignment is evident among countries regarding organisms mutated without a DNA template (often called Site-Directed Nuclease type 1 edits) as non-GMOs, whereas divergence occurs in regard to mutations introduced using a DNA template [13]. In 2018, the EU Court of Justice (EUCJ) classed all gene-edited organisms as GMOs, majorly limiting the cultivation and import of gene-edited products in the EU [15]. Following the EUCJ’s decision, the EU Council undertook a study into gene-edited organisms and published a report in 2021 which noted that the EU GMO framework was not suited to new technologies [16]. The EU initially coined these new technologies ‘New Breeding Techniques’, but now refers to them as ‘New Genomic Techniques’ and are currently undertaking further studies and consultations to possibly alter the legislation by the late 2020s [17].

In March 2023, the Genetic Technology (Precision Breeding Bill) received Royal Assent and passed into law. The Act provides primary legislation on genetic technologies and paves the way for PBOs to be grown in England and sold throughout Great Britain (GB). In relation to food safety, the bill states that PBOs must not have adverse effects on health and must not place any humans at a nutritional disadvantage. If PBOs are to form part of the food chain, there must be public confidence in the safety measures in place and trust in the regulatory processes governing the marketing of PBOs to avoid the widespread concern that arose in response to the introduction of GMOs in the late 1990s [18]. Although there is consensus within the scientific community on the safety of underlying technologies used to make PBOs, it is possible that food/feed safety may be compromised through, for example, increased toxicity, changes in allergenicity, or altered nutritional profile. PBO animals will also be assessed for risks to animal welfare. The Act does not, however, specify the precise information needed or tests to be undertaken to ensure this safety, and this will be left to secondary legislation and guidance to applicants enacted by the Department for Environment, Food and Rural Affairs (Defra), and the Foods Standards Agency (FSA). It is important to note that the Genetic Technology (Precision Breeding Bill) is an Act of the UK parliament that applies to England only and could be considered as a failure of UK-wide policy development. For Northern Ireland, EU law on GMOs continues to apply. At the time of writing, the Welsh and Scottish parliaments have not made any changes to their existing laws, so all gene-edited organisms continue to be defined as GMOs in those regions. Although the cultivation of PBOs is a devolved matter in Wales and Scotland, the UK Internal Market Act 2020 makes it clear that, if regulations on the sale of goods are complied with in one part of the UK (e.g., England), they will automatically be compliant in the other countries, making the sale of PBO food/feed cultivated in England legal in Wales and Scotland. This will inevitably lead to confusion and friction.

Applications for PBOs to be cultivated in England or imported and sold throughout GB will be first assessed by the Advisory Committee on Releases to the Environment (ACRE), who have 90 days to determine, on the basis of information supplied by the applicant, whether the product fits the definition of a PBO. If not, it could be considered as a GMO, or indeed a traditionally bred organism (TBO) (Figure 1) [19]. Although gene editing of animals is included in the PBO legislation, the ethical and welfare requirements are subject to future consideration, and the current focus is on developing secondary legislation for the marketing of new plant PBO varieties. As the Precision Breeding Bill specifies that an organism must have been able to have been created by traditional breeding, ACRE will require information on the genetic changes to determine whether an organism is a PBO. The main evidence needed is the absence of transgenic DNA and evidence that the genetic changes could have occurred spontaneously or via existing breeding methods in reasonable timeframes. In the past, ACRE has stated that they believe off-target edits produced by gene editing pose no greater safety concern than those produced by other forms of mutagenesis, and they are much rarer, so it is unlikely that there will be any undue focus on off-target mutations [20]. In addition, ACRE has previously stated that PBOs would pose “no greater risk than their traditionally bred or naturally arising counterparts”, implying that they will not consider environmental risks as part of a PBO assessment [21].

Prior to the registration and marketing of new plant varieties that incorporate genetic changes generated by new biotechnologies, such as gene editing, developers must make an application to ACRE for determination of PBO status. If it is found not to be a PBO, it will follow the existing legislation for TBOs or GMOs. Applicants who wish to commercialise food or feed products using confirmed PBOs must then apply to the FSA for marketing approval. To assess the food/feed safety of a PBO, the FSA have suggested a two-tiered approach; tier one simply requires registration of the product, and tier two requires further scrutiny to assess hazards and exposure on a case-by-case basis [22]. In deciding which tier to designate a particular PBO or derived product, the FSA will focus on the organism’s phenotype, considering novelty, composition (nutrition, toxicity or allergenicity) and other safety concerns [23]. If no hazards beyond those inherent in existing food systems are identified within these themes, the product will be assigned to tier one. If the answer to any of the triage questions is yes, then the FSA will ask for further information on a case-by-case basis. At the time of writing, this approach is yet to be piloted, and as such, there is little evidence on whether two tiers are an appropriate number, and what type of information is needed for the case-by-case safety assessment of tier-two products.

The newly proposed legislation on PBOs is distinct from, and demands less data for risk assessment, than the laws surrounding GMOs. At present, the UK’s GMO laws derive from a combination of domestic legislation, retained EU laws and obligations under the Cartagena Protocol, which are extensive and focus on the protection of human and animal health as well as the environment [24]. For a novel GMO to be approved for cultivation or for the importation of products intended for food/feed consumption, assessments are made on a case-by-case basis by Defra and the FSA for potential risks. These include assessments for toxicity, nutrition and allergenicity [25]. A public consultation is also required, and this information, as well as a risk assessment, must be published on a public register. This is reviewed by ACRE, who will advise the relevant authorities on the environmental safety of the organism. Upon authorisation, approval can last a maximum of 10 years, and during this period the producer must monitor and evaluate post-market environmental harm of the organism [24]. The authorization processes vary from region to region, but significant areas of GM crops are cultivated globally each year. In 2023, the area under GM crops was 206.3 million hectares. Twenty-seven countries cultivated a range of 11 different GM crops, with soybean the most widely planted at 100.9 million hectares, followed by maize at 69.3 million hectares, and cotton at 24.1 million hectares [26]. There are a few examples of GM animals being authorised for food, including, in the USA, AquaAdvantage salmon and GalSafe pigs. However, no GMO plants or animals are farmed commercially in the UK for food or feed; although, according to the ISAAA database, over 100 events, including those in soybean, corn, cotton and canola, have been authorised for importation, predominantly for livestock feed [27]. These foods and feeds undergo a comprehensive risk assessment and adverse health problems have not occurred due to their consumption. This is also the case in countries such as the USA, which consumes much higher levels of GMO ingredients in their diet [28]. Hence, it may be possible to use GMOs with comparable traits that have been approved as evidence of the risks of changes in gene-edited organics to human health. Indeed, there are many examples of RNAi-mediated silencing of genes controlling agronomic or nutritional traits that have been copied using gene editing [29].

In this paper, firstly, we use four case studies to explore the data requirements and risk assessment processes needed to evaluate the safety of PBOs for consumption. These case studies were chosen as viable examples of foods that producers are likely to pursue for PBO status, given their benefits to the food system and advanced stage of development. Our crop examples target a range of nutritional traits that we considered relevant to the aims of our manuscript, and which capture any potential risks or challenges for the proposed regulatory system. The PRRS pig case study was included to contrast the plant examples, and because it is relatively advanced on the pathway to commercialisation. There are also comparatively good sources of information on the genetic and biochemical changes made. We focus on the intentional changes within the PBO and how these may impact the food/feed and the consumer. Secondly, we assess whether these would be defined as PBOs and evaluate the two-tier system proposed by the FSA for novel PBOs. As PBOs are not yet marketed in the UK, these case studies are aimed to be hypothetical examples in which we can draw conclusions about future precision-bred food/feed.

2. Case Study 1—Vitamin D in Tomatoes

Enhancing the nutrient content of particular foods will enable the public to attain their recommended daily allowance (RDA) of nutrients more easily, reducing the prevalence of nutrient deficiency and related diseases and perhaps the need for vitamin/mineral supplements. Of particular benefit are healthy, inexpensive and readily available foods, such as tomatoes. Li et al. and Choi et al. have shown that it is possible to increase the vitamin D3 content in tomatoes using CRISPR/Cas9 genome editing. Vitamin D plays an important role in maintaining skeletal health, immune response and mineral homeostasis [30,31,32]. The mean daily intake of vitamin D in adults in the UK is 7.2 µg, below the RDA of 10 µg, and improving the availability of vitamin D in food could improve public health by reducing the prevalence of related diseases such as depression or cancer, particularly through the winter, where UVB radiation is low [33,34,35]. Furthermore, sales of such products could be targeted at groups particularly at risk of vitamin D deficiencies, such as children and non-white populations [36,37].

Vitamin D3 does not normally accumulate in tomato fruits; however, when the leaves are exposed to UVB, 7-dehydrocholesterol (7-DHC) is converted to vitamin D3 [38]. When not exposed to UVB, 7-DHC is converted to cholesterol via the enzyme 7-dehydrocholesterol reductase (7-DR2) and then synthesised into tomatines and esculerosides (Figure 2) [39]. A duplication of this pathway has been identified, allowing for its alteration with minimal impact on phytosterol and brassinosteroid biosynthesis [39]. Li et al. introduced a CRISPR/Cas9 construct containing two sgRNAs into Agrobacterium tumefaciens, which was then transformed into tomato plants [30]. The resulting edits blocked the activity of the SL7-DR2 enzyme, an isoform of 7-DR2, preventing 7-DHC from being converted into cholesterol (Figure 2). Subsequently, 7-DHC was converted into vitamin D when exposed to UVB.

In the T1 generation of edited plants described by Li et al., five independent knock-out alleles of the SL7-DR2 gene were identified by sequencing: three plants contained identical 108 base pair (bp) deletions, one plant had a 2 bp deletion and 1 bp insertions, and the final plant had a 1 bp insertion [30]. Genotyping and sequencing with 17 primer pairs confirmed plants in the T2 generation lacked both the Cas9 and the sgRNA expression cassettes. No off-target gene edits were discovered in the only other sterol delta reductase (S17-DR1) in these plants using PCR and sequencing, demonstrating the precision of CRISPR/Cas9. These plants would likely be classified as PBOs under the new UK and, indeed, by regulatory authorities worldwide.

Using liquid chromatography-mass spectrometry, levels of cholesterol, steroidal glycoalkaloids (tomatines and esculeosides) and 7-DHC were analysed. Interestingly, both 7-DHC and cholesterol were substantially increased in the fruits of the edited lines. In humans, when 7-DHC accumulates and results in cholesterol deficiencies, it can lead to developmental disorders; therefore, additional cholesterol may help to prevent these [40]. Furthermore, tomatoes are thought to decrease the risk of cardiovascular diseases due to phytosterols, which prevent cholesterol absorption [41]. Silaste et al. found that cholesterol levels were reduced by 5.9% in a high-tomato diet, suggesting that any risks of increased cholesterol in tomatoes are mitigated by their reductive effects [42]. Alpha-tomatine was also lower in the edited lines—a benefit to health as alkaloids can be harmful, causing gastrointestinal and neurological disorders [43]. Similarly, esculeosides A was strongly reduced in ripe fruit from edited lines; however, this may be nutritionally negative as esculeosides have anti-cancer properties and possibly inhibit atherosclerosis development [44,45]. No impact was found on phytosterol metabolism, shown through comparisons with stigmasterol of wild-type and edited lines. The consequential changes observed do not appear to pose any major health concerns.

Vitamin D in the edited tomatoes was enhanced, resulting in a concentration of 0.2 μg/g with the full RDA of an adult being obtained by the consumption of two small (25 g) tomatoes. The level of consumption could easily prevent deficiencies; however, a risk of toxicity from consuming excess vitamin D is also possible. Fortification of vitamin D has occurred safely in many products and, although rare, sporadic cases of toxicity have occurred. For example, some negative health effects were observed in children who were particularly sensitive to high vitamin D after consuming fortified milk, or from consuming products with an illegal level of vitamin D [46,47]. These cases suggest that, although uncommon, clear labelling and regulation of gene-edited products with significantly elevated vitamin levels may be beneficial in allowing the consumer to monitor their own vitamin D intake and prevent the risk of excess intake, akin to vitamin supplements. The tolerable upper limit (TUL) of vitamin D consumption recommended by the EFSA is 100 µg/d of vitamin D for adults and children above 11 years old, and 50 µg/d for children aged 1–10 years [35]. The precision-bred tomatoes described by Li et al. (2022) contain 0.2 μg/g of vitamin D3, after being irradiated with UVB for one hour. Consequently, the consumption of 500 g (approx. 20 small tomatoes per day) would reach the upper limit for those aged above 11, and half that for those aged between 1 and 10. If these tomatoes are sold only as raw products, it is unlikely that this TUL would be exceeded. Comparatively, a portion of grilled herring typically contains 19.2 µg per portion and is deemed safe for consumption [48]. Risks may occur; however, if this tomato product is processed or reduced into soups or stews, for example, the equivalent of 500 g of tomatoes could be consumed in one or two servings. People could also combine them with a diet high in foods naturally containing vitamin D, such as fish and eggs, and fortified products such as cereals and spreads.

Increasing the vitamin D content in tomatoes through gene editing is an effective way to prevent disease in deficient individuals. Consuming only two tomatoes would allow individuals to reach their RDA of vitamin D in an inexpensive and healthy way. There are still some uncertainties with this novel product, such as whether increased exposure to UVB will continue to increase the vitamin D3 content, and how tightly this will be controlled when produced commercially. Furthermore, risk managers need to consider whether this product should be freely mixed with conventional tomatoes at the point of sale or marketed through an identity-preserved distribution chain, with a clear indication of the vitamin levels to allow consumers to make informed decisions about what they are eating.

3. Case Study 2—Reduced Linoleic Acid in Cottonseed

Increasing the shelf-life of consumable products is a further application of gene editing [49]. Benefits include reductions in waste and subsequent potential reduction in land, energy and water use, with subsequent lower costs for both producers and consumers. Oils with high linoleic acid (LA), for example, are prone to oxidation, which shortens their shelf-life [50]. Oxidation may result in loss of flavour, change in nutrient content and unpleasant taste [51]. The fifth-largest source of vegetable oil consumed by humans, with high levels of LA, is cottonseed oil (Gossypium hirsutum), used in commercial and home cooking, biscuits, and margarine, among many other products [52]. Creating gene-edited cottonseed oil with reduced LA is a fast and precise method to achieve increased shelf-life.

Reducing the level of LA in cotton was achieved by Liu et al. by downregulating fatty acid desaturase (FAD2), which converts monounsaturated oleic acid into polyunsaturated LA (Figure 3.) [53]. These methods, however, use hairpin RNA-mediated gene silencing, a type of genetic modification, and as such would require pre-market authorisation as a GMO to be cultivated in the UK and elsewhere [53]. Chen et al. built upon this work, targeting the same pathway and knocking out GhFAD2-1A/D, using CRISPR/Cas9 gene editing instead [50].

Using Agrobacterium tumefaciens-mediated transformation, Chen et al. obtained CRISPR/Cas9 positive cottonseed plant transformants, identified using PCR by the presence of Cas9 and NPTII genes (from the vector pRGEB32-GhU6.9- NPT II) [50]. A mix of single nucleotide deletions and insertions occurred, along with some larger deletions of 308 and 309 bp. Three of these T0 plants were taken forward, with PCR and Sanger sequencing used to detect that intended mutations were inherited into the T1 generation; however, new mutations were also present in some plants. The authors suggest these mutations may be the result of active Cas9 or chimeric mutations in the original T0 plants. The CRISPR-P application was used to predict potential sites of off-target mutations, which were analysed using PCR and Sanger sequencing [54]. No off-target mutations were detected. Of these T1 plants, four lines were confirmed as non-transgenic by PCR and were subjected to further analysis.

Chen et al. successfully reduced the level of LA within cotton oil from 58.62% to 6.85%. As an essential fatty acid, LA is an important part of diet; however, the health benefits of high consumption are debated, with some studies suggesting that higher intake is linked to benefits such as improved cardiovascular health, and others suggesting the contrary [12,50]. Despite unclear conclusions for health, reducing the level of LA in one specific cotton variety is unlikely to cause deficiencies or significant alterations to overall LA intake, as it is readily available to consumers in many foods such as margarine, nuts and other oils [12]. Alternatively, as a longer shelf-life is an attractive trait for food producers, it is possible that many companies will increase the utilisation of this product, resulting in widespread reductions in LA availability. This is a particularly important consideration, as studies already suggest that Western populations should increase their LA intake, not decrease it [55]. The oleic acid composition of this oil was substantially increased from 13.94% to 77.72%, a similar level to that found in olive oil [56]. This increase is likely safe for consumption, as numerous studies provide evidence for the benefits of oleic acid, from immunomodulation to reduced blood pressure [57,58,59]. The percentage of palmitic acid (PA) was significantly reduced in edited lines, from 23.95% to an average of 13.18%. In line with government recommendations to reduce saturated fat intake, marketing this oil with lower PA would be beneficial for public health [60]. However, current evidence on the links between health and PA intake is inconclusive, with studies describing differing effects on diseases such as cancer and cardiovascular disease [61,62]. Any reduction of PA that may occur from cotton oil, however, could be gained from a balanced diet, as many products contain palm oil consisting of 40% PA [62].

Products with similar traits have been created using genetic modification and approved for consumption in the UK, such as soybean with reduced enzyme omega-6 desaturase, which results in a high oleic acid and reduced linoleic acid profile [63]. Hence, it may be appropriate to use evidence from the risks presented by products produced through genetic engineering when evaluating the safety of new gene-edited foods. As this product remains approved in the UK, it is likely that a comparable gene-edited product would also be safe for consumption.

Decreasing the percentage of LA in cottonseed oil via gene editing appears to be an effective way to produce a longer-lasting product. However, some variability in fatty acid profile was found under different growing conditions. For example, irrigation increased the level of LA but decreased the levels of other unsaturated fatty acids, and early planting had the opposite effect [64]. Whether the changes to the nutritional profile of the oil would have a negative or positive effect on human health are subject to debate.

4. Case Study 3—PRRSV-Resistant Pigs

Disease in livestock is a major agricultural challenge that may worsen with increasing livestock densities, changes in land use and climate change, among other factors [65,66]. Disease poses numerous problems including the risk of zoonotic crossover, poor welfare and wasted life and resources. It can also result in huge economic loss; Porcine Reproductive and Respiratory Syndrome Virus (PRRSV) results in a £26 m loss in the UK alone through abortions, stillbirths, and disease control [67]. To tackle these problems, Burkard et al. have demonstrated the possibility of generating gene-edited pigs with resistance to the PRRS virus [68].

Resistance was achieved using CRISPR/Cas9 to delete part of the CD163 gene, which codes for the extracellular scavenger receptor cysteine-rich (SRCR) domains that act as a receptor for PRRVS. The CD163 protein, which is present in specific macrophages, has nine SRCR domains that sit extracellularly (Figure 4). This protein is known to have many biological functions; hence, knocking out the entire gene could have negative effects on the animal’s health, in turn affecting meat quality [69]. The SRCR5 domain, however, is not known to have any specific roles; therefore, it was chosen for knock-out. A pair of guide RNAs was used to target the intronic regions that flank exon 7 of the CD163 gene, which encodes SRCR5.

The Cas9 mRNA and sgRNAs were injected directly into zygotes to knock out the SRCR5 domain. Live piglets were born with the intended deletion at exon 7, revealed by genotyping via PCR and Sanger sequencing. Of these piglets, three also had insertions of new DNA at the targeted site, resulting from non-homologous end-joining repair. One pig homozygous for the deletion and one heterozygous pig were mated, producing six heterozygous, two homozygous, and two wild-type offspring. Cells from the two homozygous T1 pigs were tested in vitro and were found to be highly resistant to infection of PRRSV. Furthermore, this process was repeated in vivo by Burkard et al., showing that SRCR5 knock-out pigs are completely resistant to even highly virulent PRRSV-1 subtype 2 strains [70].

No adverse effects in the pigs were detected from the deletion of SRCR5; they had normal growth rates and standard blood counts [68]. CD163 still exhibited correct folding and displayed its function as a haemoglobin-haptoglobin scavenger, with tests showing that haemoglobin levels in the blood remained the same in wild-type and edited lines. The levels of mRNA expression of CD163 and protein were equivalent in edited lines, suggesting that the nutrient level of the pigs remained unchanged. The above factors indicate that pigs remained in full health despite the knock-out, indicating that these gene-edited pigs are likely safe to eat. With the main biological functions of the CD163 relating to the immune response, it is unlikely that the deletion of the SRCR5 domain will have any changes in terms of the nutritional content [71].

PRRSV-resistant pigs are a success story of CRISPR/Cas9 and demonstrate its use in animal health and welfare, reducing waste and improving economic gain. Although Burkard et al. do not mention whether checks were undertaken to determine if off-target mutations were made, these would be needed before PBO status could be established [70]. There is nothing obvious to indicate that the loss of the SRCR5 domain would alter the welfare or the safety of food/feed products from these animals. However, additional work on elite breeding animals would be needed to confirm this. As there is no intention to alter the nutritional profile of the consumed products, an identity-preserved distribution chain to consumers would not be necessary.

5. Case Study 4—Reduced Asparagine Wheat

Gene editing has been used to reduce harmful components of foods such as acrylamide, a carcinogenic substance. Acrylamide forms from a reaction between amino acids and sugars, mainly via the Maillard reaction which ‘browns’ and alters the taste of foods during heating (Figure 5) [72]. The level of acrylamide depends mostly on the asparagine and sugar content of the food, and the cooking temperature and duration, but it is also linked to the cooking method, pH, moisture content and storage [73]. Foods that contain precursors of acrylamide include pizza, bread, cakes, and cereals, many of which are produced from wheat [74]. As wheat is so widely consumed, reducing the formation of acrylamide in foods could significantly reduce the public health risk of this carcinogen.

Raffan et al. used CRISPR/Cas9 genome editing in wheat (Triticum aestivum) to knock out a gene responsible for asparagine synthetases; enzymes which catalyse the conversion of glutamine and aspartate to glutamate and asparagine [75]. Wheat has five asparagine synthetase genes per genome; however, the authors focused on TaASN2 as it is highly expressed in the seeds, the edible part of a wheat plant. Four sgRNAs were designed to target the first exon of the three homologues of the TaASN2 gene and were introduced into wheat plants along with the Cas9 enzyme.

Successful transformation was confirmed and T0 plants were self-pollinated to produce T1 seeds, which were germinated, grown to maturity, and analysed. Next-generation sequencing detected heritable deletions, single bp insertions and substitutions. Homozygous knock-out plants were produced and transgene-free plant lines with reduced asparagine content were selected.

Analysis of edited seeds showed that most of the TaASN2 protein was truncated and non-functional. In some lines, the levels of free asparagine were significantly lower than controls and the acrylamide levels in heated flour samples made from the grain of gene-edited lines were reduced by 44–45% compared to the wild-type.

Current methods to remove or reduce the formation of acrylamide pose challenges, such as altering mouthfeel and other sensory perceptions of the food or needing to combine many removal methods to be effective [76]. Acrylamide-forming potential also varies between wheat varieties and growing conditions; it increases, for example, in response to sulphur deficiency and pathogen infection (reviewed by Raffan & Halford [77]). A better understanding of the environmental factors, combined with low acrylamide genotypes such as those described by Raffan et al. could markedly reduce the acrylamide in cooked bakery products [75]. Although not a food safety issue, Rafan et al. report that low asparagine seeds showed poor germination, but that this could be overcome by exogenous application of asparagine [75].

6. Discussion

The gene-edited organisms described above demonstrate the precise and time-efficient capacity of this technology to produce foods with higher nutritional quality, longer shelf-life, disease resistance and reduced harmful substances. At present, a PBO in the UK is defined by legislation as a product of modern biotechnology that could have occurred through traditional breeding. Although we have little experience of precisely which products of modern biotechnology ACRE will define as PBOs and which it will decide are not, we believe that all our case studies would be defined as PBOs.

It is easy to see how individual edits, such as those exemplified above, could be defined as PBOs. However, it is less clear how multiple edits in the same organism would be processed. Theoretically, given enough time, a vast range of traits could occur via conventional breeding. Seven thousand years of domesticating bananas, for example, has created a seedless, parthenocarpic fruit barely resembling its ancestor [78], yet this level of change in one PBO application (i.e., the one-step domestication of banana) would probably be beyond the definition as per the Act. It is likely that organisms with insertions, deletions and substitutions resulting from editing would all be considered PBOs, as it is obvious that these are utilised in conventional breeding [79]. However, the vast majority of deletions, substitutions and insertions are small (say, less than 1000 bp), and it will be interesting to see how ACRE conclude on chromosome-scale edits or multiple edits that would have taken very long timescales to achieve in other ways [80].

Within the FSAs triage questions, the phrase ‘a significant change to nutrient quality’ is open to wide interpretation as to what will be deemed a significant change. According to Entine et al., the model used by Health Canada refers to whether changes in the expression of a trait are 25–30% higher or lower than the wild-type to define a novel product and therefore require a pre-market risk assessment [81]. The UK FSA has not stated the level of change to nutrient quality that would require a tier two assessment, but it is likely to be a case-by-case judgement. For instance, assuming other nutritional components of the product were equivalent, reducing a potentially unhealthy component, such as acrylamide, in wheat via a knock-out mutation, is unlikely to need to go through the higher tier of risk assessment. In contrast, malnutrition may theoretically occur from the changes observed in case studies 1 and 2 due to possible overconsumption of vitamin D and a reduction in the availability of LA, respectively, indicating that these could be flagged for specific risk management such as an identity-preserved distribution chain. In the current model, PRRSV-resistant pigs are likely to pass through as tier one, as the target gene of the PRRSV pigs case study is mostly associated with immune system roles and unlikely to affect the meat and therefore would not be defined as tier two, based on nutrient alteration.

The variability of the altered components in PBOs may pose a challenge to the FSA. For example, case study 4 reported a range of free asparagine levels, from 9% to 56% of the wild-type levels, which was then further altered by heat stress or soil sulphur levels. The level of vitamin D in ripe tomato fruits was also reasonably variable from 0.174 and 0.255 μg/g [30]. In contrast, case study 2 had a <10% range on the levels of linoleic and oleic acid in seeds of the plant lines tested. However, the examples described above are from research trials, not the analysis of new varieties grown commercially. It also must be seen in context to the variability known to occur naturally and currently managed or accepted in the food chain. Much of this variability originates from where and how an organism was grown [82], storage conditions [83], and other phases of the food’s lifecycle. For example, vitamin C in conventionally bred broccoli grown in autumn can be almost twice as high as broccoli grown in the spring [84], and strawberries grown with compost have significantly higher levels of potassium, but lower levels of iron [85]. In case study 1, Li et al. only exposed the fruits to one hour of UVB in controlled lab conditions, and it would be informative to understand what the levels of vitamin D would reach when grown in conventional conditions, or in different locations with variable UVB exposure [30].

It is clear that whatever the change in nutrient quality, a PBO will not be considered under the retained Novel Foods Regulation (EU) 2015/2283 (the Novel Food (Amendment) (EU exit) regulations 2019 No. 702) unless the animal or crop species that was edited complies with the inherent definitions of a novel food [86]. However, if an application is made for an organism that is both a PBO and a novel food according to (EU) 2015/2283, then a risk assessment would need to include the all necessary aspects, including the identity of the novel food, production process, compositional data, specifications, the history of use of the novel food and/or of its source, proposed uses and use levels and anticipated intake, absorption, distribution, metabolism and excretion, nutritional information, toxicological information and allergenicity [87].

Many PBOs may be safe when considered in isolation, but hazards may become evident when too many products with the same trait are released into the market. Li et al. point to the possibility of increasing vitamin D in many other crops from the Solanaceae family, such as peppers, potatoes and aubergine [30]. However, this may result in widespread vitamin D availability and could increase the risk of toxicity. Similarly, LA has already been decreased by gene editing in high-oleic acid soybean, which has been released into the US market [88]. These examples demonstrate the likelihood of multiple products being edited for the same trait and reveal the possibility for these widely used products to cause malnutrition. To avoid this, additional measures may be required by risk managers to maintain a range of available food products combined with consumer information on healthy and appropriate diets.

The off-target effects of gene editing have attracted attention; however, ACRE has stated there is no scientific evidence to suggest that they pose any risk greater than in conventional breeding [20]. Since the publication of this advice, further evidence has emerged on the nature of off-targets; a review by Sturme et al. backs up ACRE’s view, finding that in gene-edited plants, the mutation frequency is lower than in conventionally bred plants, and in general, the off-target mutations are small insertions or deletions [89]. Obviously, there is still a small risk of unintended genetic changes resulting from editing but, if deleterious, these would be selected against during the breeding steps resulting in new varieties [90].

Checking for the presence of transgenic DNA is a key component in confirming an organism is a PBO, and in our examples, PCR was commonly undertaken to screen for transgene-free individuals. Case studies 1 and 4 required breeding until generation T2 for transgene-free organisms, whereas case study 2 achieved this by generation T1. The number of generations is likely to vary on a case-by-case basis depending on crop species and the number of crossed or selfed individuals taken forward. However, our case studies come from research publications, not PBO applications, and regardless of how it is achieved, ACRE makes it clear that if a transgene is used to achieve editing, it must be removed, for example, by segregation.

6.1. Traceability and Labelling

Defra and the FSA have stated that they will develop public registers of PBOs used for food or feed. In addition, the British Society of Plant Breeders has committed to maintaining a public register of precision-bred crop varieties to support transparency and openness of information within the supply chain [91]. Although the public appears to be more accepting of PBOs than GMOs [92], 84% of people still believe the labelling of products is important [93]. The Precision Breeding Bill has no mention of on-pack labelling, although it does state that the government may ‘impose requirements for the purpose of securing traceability’, enabling PBOs to be detected and differentiated from traditionally bred varieties, and subsequently labelled. Tracing of PBOs may be problematic because identical mutations can occur spontaneously; however, detection methods for specific examples of PBOs have been developed [94,95]. A further possible difficulty with individual labelling may come down to space on packaging, especially when surveys have found that consumers think abbreviations such as ‘GE’ are not sufficient [93]. As part of a pre-market risk assessment, the FSA could advise risk managers regarding labelling conditions on a case-by-case basis to ensure the safety of particular consumer groups. For example, in case study 1, the tomato vitamin D content is significantly higher than would normally be expected in this fruit; therefore, an identity-preserved distribution and clear labelling at the point of sale might be recommended to enable customers to monitor their own level of intake [96].

6.2. Global Outlook

There is little formal, global harmonisation of regulatory frameworks applied to the products of gene editing. For example, different countries have adopted different definitions and even different names for the types of gene-edited organisms they exclude from their GMO legislation. This will inevitably lead to confusion and additional challenges to food producers looking to trade, even between countries where gene-edited organisms are not classed as GMOs. Additionally, the expensive and time-consuming approval process required for GMOs is likely to prevent PBO producers from trading these products with countries such as New Zealand, the EU, Peru or South Africa, where gene-edited products are currently regulated as GMOs [97]. Some countries have adopted regulations that treat gene-edited animals and plants similarly, whereas others, such as the US and Chile, regulate the two very differently [98]. This heterogeneity in definitions and legislative procedures, alongside the difficulties in uniquely identifying products of gene editing, will pose significant challenges to international trade and compliance.

Although the legal frameworks adopted by different countries differ in their detail, many are coalescing around some core principles similar to those described here for the UK. It is hoped that agreements over equivalence can be reached where gene-edited products with no transgenic DNA are not classed as GMOs [13]. In Central and South America, multi-national agreements have been created to improve harmonisation and reduce inconsistencies in regulations [11,97]. In 2018, 13 nations across the Americas issued a joint statement to the World Trade Organisation in support of relaxed regulations for gene-edited crops, and in 2019 Honduras, El Salvador and Guatemala signed agreements to harmonise the commercialisation of crops [98]. It is likely that other countries highly interconnected by trade will follow suit. England is on a path to include gene-edited animals in new regulations, and trade opportunities may arise from treating gene-edited animals who meet the same animal welfare requirements as conventionally bred animals, in line with many other countries including Japan, Brazil and Canada [98]. Countries heavily reliant on genetically modified imported food and feed may be influenced into adapting legislation to allow for products of gene editing if they become as widespread as GMOs. Turnbull suggests that countries leading in the trade of genetically modified foods are ones which are able to quickly adapt laws and accommodate gene-edited products [11]. A lighter regulatory approach will also lead to greater numbers of commercial applications for gene-edited products from small enterprises and research institutes, as seen in Argentina [13].

Even if harmonisation over definitions is reached, approval of products for trade may take time, as individual countries need to ensure these products meet their national food safety guidelines. These differences may result in the authorisation for marketing in one country not automatically translating to others and additional checks may be needed unless greater regulatory harmonisation is reached.

6.3. Consumer Understanding of PBOs

To utilise the potential of precision breeding for crops and farmed animals, public acceptance and trust in the safety of these products must be achieved. A report commissioned by Food Standards Scotland found the vast majority of participants started from a low base of awareness but were open to learning more about the science of how food crops were developed [99]. In addition, the FSA’s ‘consumer perceptions of gene edited foods report’ found that people who felt more informed about the topic were more likely to believe that these foods should be sold in the UK, suggesting that greater awareness and understanding may help to drive the success of PBOs [93]. They also found that, in general, people’s knowledge of gene editing was lacking, and therefore it may be beneficial for a governmental body such as the FSA to create clear and accessible information for the public on PBOs through websites, television and social media. Campaigns promoting the range of benefits from PBOs may reach a wider audience, in particular, promoting the role in animal welfare has been found impactful [100]. Furthermore, interchanging the use of words such ‘gene editing’, ‘genome editing’ or ‘genome engineering’ has been found to add confusion in educating the public [101]. Similarly, interchanging the terms ‘gene-edited’ and ‘PBO’ will cause confusion. It should be clear that gene editing can be used to produce a PBO but that not all products of gene editing are necessarily PBOs. Enhancing knowledge and trust may also reduce the desire for all PBOs to be labelled at the point of sale, reducing costs associated with packaging.

7. Conclusions

Alongside non-biotechnological methods, PBOs offer the potential to tackle numerous agricultural and food system challenges. The Precision Breeding Act 2023 is a major change in the regulatory landscape for genetic technologies in the UK. Our four case studies demonstrate the possible risks and benefits of PBOs, and we discuss how, if they progress to market, they may be managed within the proposed two-tier system. The UK’s definition of a PBO is in many respects harmonious with other countries, but it remains to be seen whether, in practice, it will facilitate seamless trade of seeds and food/feed products into and out of the UK. Consumer surveys reveal considerable confusion and uncertainty over the differences between various biotechnologies, which could result in barriers to the commercialisation of PBOs. To mitigate this, organisations that engender public trust, such as the FSA, might consider developing policies to improve consumer awareness of the technologies and confidence in the safety assessment of food/feed derived from PBOs.